Generally speaking, it is challenging to correlate changes in bicycle ridership and safety at a city-wide level to a particular program or project such as the Bicycle Boulevard Network. Many factors can affect bicycle usage and safety, including economic trends such as the price of gasoline or variations in employment levels, overall transportation activity, and other government efforts such as public safety education programs. And while planners and engineers utilize available data in making decisions about bicycle infrastructure, they frequently must also rely on the feedback of users (cyclists) and accepted engineering best practices in lieu of a complete dataset. Berkeley is no exception in this regard.
While it is difficult to establish a relationship of causality or correlation between Berkeley’s Bicycle Boulevard Network and changes in usage and safety, we can still speak meaningfully about usage and safety trends that have taken place concurrently with the development of the network. Subjective safety measures that have to do with cyclists’ perception of the cycling environment—such as rider route choice and other ridership trends—demonstrate that Bicycle Boulevards are desirable for cyclists and may encourage more bicycle use. Objective safety measures that have to do with available data such as changes in cyclist volumes and collisions demonstrate that as volumes have increased collisions have decreased. This suggests that the provision of facilities that are desirable to cyclists is in itself a safety strategy, insofar as it results in a safety in numbers effect.
Subjective Safety Evaluation
The Milvia Slow Street project—a progenitor of the modern Bicycle Boulevard—is an example of the effect that desirable infrastructure can have on route choice. Following the implementation of the Slow Street, one location observed volumes of cyclists during the p.m. peak commute hours increased by 117 percent and at another by 49 percent. A small portion of this increase might be accounted for by natural daily variation in bicycle travel, though this still leaves a significant change in bicycle volumes unexplained. It is reasonable to assume that a massive and sudden increase in overall bicycle activity did not occur during this time period in Berkeley, which means that cyclists must have been diverting from other, less inviting parallel routes such as Shattuck Avenue and Martin Luther King Jr. Way, and choosing to ride on the slower, calmer Milvia Street. This suggests that the character of the
roadway is a critical element in route choice and by extension, in the decision of whether or not to bicycle at all. Data from at least one study confirms this relationship between Bicycle Boulevards and levels of cycling, particularly among less experienced cyclists. Research indicates that bicycle boulevards may be more effective than bike lanes on arterials at encouraging more bicycling among groups of people who currently do not bicycle much.288
These conclusions confirm the results of historic and contemporary surveys of cyclists in which comments are focused on the dangers of sharing the roadway with high volumes of fast-moving automobiles.
The City of Berkeley has conducted manual bicycle counts at approximately10 locations throughout the Bicycle Boulevard network on an annual basis since 2000. These counts, which take place in late September and early October, capture the number of cyclists during the two peak hours from 4-6 pm. The counts also capture observed gender, sidewalk riding, and helmet use. The city is currently analyzing the 10-year bike count data in order to prepare a public report of bicycling trends in Berkeley. The analysis provided below was conducted on the 2000-2009 data as part of the development of that draft report.
One of the goals of establishing on-street bikeways is the encouragement of cyclist compliance with traffic laws, such as riding in the correct direction on the roadway. If bikeways are successful in this regard, such behaviors should be reduced, producing a potential decrease in the types of collisions caused by these illegal behaviors. Data reflecting the number of cyclists riding on sidewalks was only recorded at three intersections— Bowditch & Channing, California & Russell and MLK & Russell—and only for the years 2000, 2003, 2005 and 2009. During the years it was observed, sidewalk riding decreased modestly on Bowditch & Channing, from 5 percent in 2000 to 3 percent in 2009. A larger decrease was observed at California & Russell from 10 percent in 2000 to 5 percent in 2009. MLK & Russell experienced a large decrease in sidewalk riding from 2000 to 2003, decreasing from 12 percent to 5 percent. Given that one of the design goals of the Bicycle Boulevard pavement legend is to identify the treated roadway as a bicycle priority street, it may be reasonable to assume a connection between these treatments and sidewalk riding behavior. During the years 2000 to 2002, the city was in the process of installing Bicycle Boulevard signs and markings throughout the network. This time period corresponds to the dramatic decrease in sidewalk riding at the intersection of MLK and Russell between 2000 and 2003.
At least one study has shown that women cyclists serve as “indicator species” for bike friendliness because women are more likely to go out of their way to use bike boulevards. In a study conducted by Portland State University, the stated and revealed preference data comparing men and women found that women are more likely to prefer to bicycle on low-traffic streets and bicycle boulevards, and less likely to prefer riding on busier streets with bike lanes. Similarly, less experienced bicyclists placed higher importance on factors that make the trip easier—routes with less traffic and requiring less physical effort.289 In Berkeley, the gender ratio of cyclists using the Bicycle Boulevard system has
become more balanced since the year 2000 when the male-female split was 70/30. In 2009, the male-female ratio observed was 60/40. This suggests that women have become more comfortable riding bicycles in Berkeley on the Bicycle Boulevard network and are accounting for a larger portion of the total pool of cyclists observed, in turn contributing
to the citywide increase for cyclists. According to the “indicator species” theory, this data also suggests that Bicycle Boulevards have been more effective at inviting both women and less experienced cyclists to use the network and contributing to the safety in numbers effect.
Objective Safety Evaluation
In an ideal situation, a data-driven analysis of the effectiveness of the Bicycle Boulevards at promoting safety would involve intersection-level analysis of bicycle volumes and collisions. This analysis would identify specific ways in which the Bicycle Boulevard treatments have served as a countermeasure to particular Primary Collision Factors (PCFs). However, there is not sufficient data at the intersection level to conduct such an analysis. As an alternative approach, the city is analyzing trends at a citywide level in order to identify network-level changes in absolute numbers of bicycle volumes and collisions and to identify—in a preliminary fashion—trends in the rate of collisions on the network. Note that given the small sample size, none of the trends identified below are considered statistically significant, pending further analysis.
Berkeley’s Bicycle Boulevard network counts show a 72 percent increase in levels of cycling from the year 2000 to 2010. Bicycling increased on average about 5 percent per year during that time period. The single biggest increase in bicycle usage on the network occurred between 2002 and 2003 when cycling rose by 11 percent within one year. The time period of September 2002-September 2003 corresponds with the installation of Bicycle Boulevard signs and markings through the city’s bike network.
Collisions at the bicycle count locations decreased by just over 5.5 percent during the same time period, 2000-2009, when cycling activity had increased 46 percent. From 2002 to 2003, following implementation of Bike Boulevard signs and markings and during the single biggest expansion of cycling activity, collisions decreased approximately 5.3 percent. Bicycle volumes are collected only once per year, whereas collisions are compiled from state records from an entire year. For this reason, it is impossible to create a true rate of collisions per cyclist—the data are not complimentary. Keeping this in mind, the city created a safety factor, which is a ratio of available data (ratio of cyclist volumes to total collisions). The volumes to collisions factor decreased overall by 35.4 percent from 2000- 2009, an average of 6.6 percent per year.
While the data and analysis provided above are preliminary and subject to further refinement, the trends are fairly clear. For the period of time from 2000-2009, absolute numbers of bicycles increased while absolute numbers of collisions decreased. Despite the fact that there were more cyclists on the road, numbers of collisions did not increase. Of particular interest is the trend during the years 2002-2003, when the Bicycle Boulevards were first implemented. From a before and after perspective, the dramatic change in absolute numbers of cyclists, and corresponding decrease in collisions, is compelling. Again, while it is difficult to establish any firm causality relationship, this correspondence is unlikely to be a mere coincidence.
CONCLUSIONS
Berkeley’s experience implementing its Bicycle Boulevard network provides a number of lessons learned for communities wishing to undertake a similar project. No doubt Berkeley’s success from a subjective safety perspective is largely attributable to the way the infrastructure safety strategy was based directly on cyclist preferences. Surveys dating back to the 1970s in Berkeley and elsewhere indicate cyclists’ preference to use a facility separated from high-volume, high-speed automobile traffic. The Bicycle Boulevards, through their opportunistic use of existing traffic-calmed streets from the 1970 and new traffic calming measures, created an environment that matched cyclists’ preferences. Bicycle volume data supports, at least in part, the success of this strategy at encouraging cycling. This further emphasizes the need for robust public outreach processes to assess cyclists’ specific needs when planning effective bikeway networks.
As has already been stated, there are serious difficulties in establishing a causal relationship between the establishment of bicycle boulevards and improvements in objective safety. However, it is difficult to discount the compelling increase in cycling and corresponding decrease in collisions that occurred during the period of the available data, especially during the years immediately before and after the signing and marking of the network. While the trends suggest promising outcomes for future Bicycle Boulevard improvements, additional data collection and study is needed to establish a statistical relationship of causality. For example, a more robust, year-round count program would correspond to the year-round collision data available from the state. Intersection level data and observations should be conducted to identify whether or not the characteristics of Bicycle Boulevards serve as effective countermeasures for specific Primary Collision Factors. In addition to collecting more robust volume and collision data at the intersection level on the Bike Boulevard network, such data should be collected on comparable segments and intersections off the network in order to provide a control for ongoing trend analysis.
A timeline of the significant milestones discussed above follows:
1968 The Berkeley General Plan set policy discouraging auto traffic, calling for “…a fully integrated system of pedestrian, bicycle, local transit…”
1970 The Berkeley Planning Commission conducted a survey of local bicyclists to determine what factors prevented people from riding bicycles.
1971 Berkeley Bikeways Plan among the first attempts in the nation to outline a strategy for encouraging bicycle use; first Berkeley bicycle network.
1974 Berkeley Neighborhood Traffic Study led to first traffic diverters; lays ground- work for future traffic-calmed Bicycle Boulevard Network.
1975 41 new traffic diverters were installed in the city, as well as 18 traffic circles, in an effort to reduce traffic impacts on residential neighborhoods.
1975 “Citizens Against the Barricades” had formed and began collecting signatures for a ballot proposition to remove the new traffic devices.
1976 Between 1976 and 1980 propositions against barricades were defeated, but proponents won in the Supreme Court. Then decision was rendered moot by legislation.
network and policies established in the 1971 Berkeley Bikeways Plan.
1999 Berkeley held a series of public workshops to finalize the conceptual design of the Bicycle Boulevards, including signs, markings and traffic calming.
2000 Berkeley Bicycle Plan adopted by Berkeley City Council, including the Bicycle Boulevard Design Tools and Guidelines document.
2000 Berkeley City Council authorized staff to implement the bicycle boulevards, us- ing the Design Tools and Guidelines Report as a guiding document.
2000 The City began constructing Bicycle Boulevards along opportunity corridors created by previous traffic calming efforts and older bicycle routes.
2001 First signs and pavement legends installed on Hillegass/Bowditch Bicycle Bou- levard for demonstration/evaluation.
2002 Berkeley I-80 Bicycle Pedestrian Bridge constructed connecting Berkeley’s neighborhood bikeways to the Marina and Bay Trail over Interstate 80.
2002 Traffic signals installed at Russell and Telegraph and Virginia and 6th Street to improve crossing for bicyclists; Russell & Piedmont diverter upgraded.
2003 Signs and pavement legends installed on remaining six boulevards, complet- ing the first phase of Bicycle Boulevard implementation.
2005 Berkeley Bicycle Plan updated to be largely consistent with contemporary Cal- trans design standards.
2009 The gender ratio of cyclists using Bicycle Boulevard network is approximately 60/40 (male/female), up from 70/30 (male/female) in 2000.
2010 Counts show a 72 percent increase in cycling between 2000-2010. Collisions at the bicycle count locations decreased by just over 5.5 percent between 2000 and 2009.
Maps describing the Berkeley Bicycle Plan, the Berkeley Bikeway Network, and Berkeley Bicycle Boulevard Network follow:
Figure 15. Berkeley Bicycle Plan, the Berkeley Bikeway Network, and Berkeley Bicycle Boulevard Network